1104
Characteristics of Hepatitis B X Antigen, Antibodies to X Antigen, and Antibodies to the Viral Polymerase during Hepatitis B Virus Infection Norio Horiike,* Baruch S. Blumberg, and Mark A. Feitelson*
Fox Chase Cancer Center, Philadelphia. Pennsylvania
Hepatitis B virus (HBV) infection is worldwide. Its pathogenesis is variable in that some individuals develop acute hepatitis while others develop the often deadly fulminant hepatitis and still others develop the asymptomatic chronic carrier state. Long-term infections could lead to the appearance of mild chronic persistent hepatitis or more severe chronic active hepatitis. Carriers with chronic active hepatitis are at high risk for the development of liver cirrhosis and primary hepatocellular carcinoma [I. 2]. The viral serum and liver markers associated with each of these disease states have been the subject of extensive investigations, but much remains to be done before a thorough understanding of their roles in the pathogenesis of liver disease is attained. The genome of HBV consists of four open-reading frames, each of which encodes one or more viral gene products. The first encodes a family of related hepatitis B surface antigen (HBsAg) polypeptides in the envelope of viral and subviral particles found in infectious serum. The second encodes the hepatitis B core antigen (HBcAg) polypeptides of the viral nucleocapsid. The third encodes products with reverse transcriptase, ribonuclease H, and DNA polymerase activities
Received 25 February 1991; revised 28 May 1991. Presented in part: Meeting on the Molecular Biology of Hepatitis B Viruses, University of California, San Diego, August 1990. Grant support: National Institutes of Health (RR-05895 and CA-40737, -06927, and -48656), Johnson and Johnson (Focused Giving grant 830702), Commonwealth of Pennsylvania, and Ehime University School of Medicine (to N.H.). Reprints or correspondence: Dr. Mark Feitelson, Department of Pathology and Cell Biology, Thomas Jefferson University School of Medicine. 1020 Locust sr., Philadelphia, PA 19107. * Present addresses: Third Department oflnternal Medicine, Ehime University School of Medicine. Ehime-ken, Japan (N.H.); Department of Pathology and Cell Biology, Thomas Jefferson University School of Medicine. Philadelphia (M.A. F.). The Journal ofInfectious Diseases 1991;164:1104-12 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6406-0009$01.00
required for viral replication. The fourth encodes hepatitis B X antigen (HBxAg) polypeptides [1, 3]. Recent work in renal dialysis patients who became transiently or persistently positive for HBsAg has shown that antibody directed against the viral polymerase (anti-pol) is the earliest marker of HBV infection [4], that HBxAg in the serum correlates with other markers of viral replication [5, 6], and that the appearance of antibodies against HBxAg (anti-HBx) correlates with a decrease in the amount ofdetectable virus in the serum [7]. The finding of fusion polypeptides with HBxAg plus core antigen determinants in viral nucleocapsid particles [8], in which viral replication is known to take place [9], further supports the inference that HBxAg is involved in replication. The finding of both HBxAg and polymerase polypeptides associated with viral replication suggests that they may be targets of antiviral immune responses likely to be important to the outcome of infection. This is especially important since liver disease is thought to be due to host immune responses against hepatocytes expressing one or more HBV antigens and not to a direct cytopathic effect of HBV [10]. If so, then the patterns of HBxAg [5, 6, 11-14], antibody to X protein [7, II, 1519], and anti-pol [4, 20, 21], which have only recently come under intense study, may be important markers of the hostvirus relationship. In this study we determined the presence, frequency, and distribution of these newly characterized HBV markers in sera and livers from Japanese patients infected with HBV. The results were compared with the clinical diagnosis to determine the role of these markers in the natural history of infection.
Materials and Methods Patients. Eighty-five adult HBV-infected Japanese patients recruited from the hospital at the Ehime University School of
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
The characteristics of hepatitis B virus (HBV) X antigen (HBxAg) and antibodies against the X antigen (anti-HBx) and the viral polymerase (anti-pol) were determined in 85 HBV-infected patients. HBxAg was detected in sera positive for HBV e antigen (HBeAg) and HBV DNA in patients with acute and chronic hepatitis, while anti-HBx appeared when markers of viral replication became undetectable. HBxAg was common in the liver among patients with chronic hepatitis independent of HBV replication markers but was closely correlated with elevated alanine aminotransferase, implying that HBxAg in liver may be important in the pathogenesis of chronic infection. Anti-pol was detected in many samples positive for HBeAg and HBV DNA and less often in serum samples without markers of HBV replication, suggesting that this marker could reflect ongoing viral replication in the liver, even though such markers were absent from sera.
110 1991; 164 (December)
New Markers in HBV Infection
the anti-HBx peptide antisera [6]. Recombinant HBxAg polypeptide in this case was made from the construct pET8c-X3 expressed in BL21(DE3)pLysS as described earlier [24], which was a gift ofW. Robinson (Stanford University, Stanford, CA). Specificity in the anti-pol solid-phase assay was verified by determining whether test sera could immunoprecipitate a recombinant DNA-derived polymerase polypeptide fragment. The immunoprecipitation was carried out exactly as described for anti-HBx above, except that the material for immunoprecipitation was derived from in vitro transcription and translation of a polymerase expression vector. The polymerase fragment chosen for expression included the 3' end of the polymerase gene from the full-length clone of HBV adw, [25], which was amplified using primers spanning nucleotide positions 2436-2412 (MF04) and 631-660 (MFHP 2 ) . These primers were made in the oligonucleotide synthesis facility at the Fox Chase Cancer Center. This region of the polymerase gene was chosen for expression because it carried imrnunodominant determinants recognized by most patients with an anti-pol response [4, 26]. The expression product (32 kDa) was identified by irnrnunoprecipitation using appropriate anti-pol peptide antisera [4]. Tritiumlabeled polymerase polypeptide was then used for immunoprecipitation followed by SDS-PAGE, and the results were visualized by fluorography [27]. Antibodies against hepatitis delta virus (anti-HD) were measured using a commercially available assay (Abbott). Serum alanine aminotransferase levels were measured by standard techniques [4]. Peptide synthesis and generation of corresponding antisera. Synthetic pep tides from the polymerase and X regions of HBV DNA (subtype ayw) [28] were made at the solid-phase peptide synthesis facility at the Fox Chase Cancer Center. Corresponding antisera were prepared as described [4, 5, 7]. Immunohistochemical study. The liver specimens were fixed in formalin and embedded in paraffin. Sections (6 JLm thick) were used for staining as described [22]. HBxAg was examined by an avidin-biotin complex method using a polyclonal rabbit antibody to an HBxAg-specific peptide spanning residues 100114 (anti-99) [7]. Briefly, slides were blocked with excess avidin (0.05%, 10 min) and biotin (0.05%, 10 min) and then incubated overnight with a 1:400 dilution ofanti-99. After washing, slides were incubated with biotinylated goat anti-rabbit immunoglobulin for 30 min, then with avidin-biotinylated peroxidase complex (Vectastain, ABC kit; Vector Laboratories, Burlingame, CA), and finally with diarninobenzidine substrate [22]. Slides were counterstained with hematoxylin and observed by light microscopy under code. Some slides were stained with a primary antibody raised against full-length, 17-kDa HBxAg polypeptide, made by recombinant DNA techniques and provided by W. Robinson. HBsAg and HBcAg were detected by using commercially available kits (Dako, Copenhagen) following the instructions of the manufacturer. HBxAg expression in the liver was graded as follows: -, negative; +, positive, with a spotty distribution in tissue; ++, positive, with a lobular or diffuse distribution in tissue. The specificity of staining for HBxAg was demonstrated by replacement of anti-99 with preimmune rabbit serum or by blocking the stain-
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
Medicine were studied. Among them, 4 had fulminant hepatitis, 7 had acute hepatitis, 10 were asymptomatic carriers (defined as persistently HBsAg+ patients lacking clinical and biochemical evidence of liver disease), 31 had chronic hepatitis, 14 had cirrhosis, and 19 had hepatocellular carcinoma. For patients from whom liver biopsy materials were available (3 with fulminant hepatitis, 2 with acute hepatitis, 14 with chronic hepatitis, 3 with cirrhosis, and 7 with hepatocellular carcinoma), paraffin-embedded sections were stained for HBsAg, HBcAg, and HBxAg as described [22]. HBV and hepatitis delta virus (HDV) markers and alanine aminotransferase levels in serum. The presence of serum HBsAg, antibodies against HBsAg (anti-Hlss), HBV e antigen (HBeAg), antibodies against HBeAg (anti-Hlse), and antibodies against HBcAg were determined using commercial radioimmunoassays (Abbott Laboratories, North Chicago, IL). HBV DNA was tested by dot hybridization using 1 JLl of each serum sample as described [5]. HBxAg was determined exactly as described [5]. Briefly, serum samples were initially screened for HBxAg by ELISA. Microtiter wells were coated with human serum for testing, then with anti-HBx made in rabbits against HBxAg synthetic peptides; finally, an enzyme-conjugated anti-immunoglobulin and substrate for color development were added. Positive values were scored as being >2 SD above the mean of the negative controls, which was a panel of human serum samples from uninfected individuals. The specificity of binding was determined by repeating the assay with normal rabbit serum in the place of peptide antibodies. Specificity was also shown by preincubation of the peptide antibodies with the immunizing peptides before assay. Specificity of the solid-phase assay was verified by immunoprecipitation of selected positive and negative serum samples using a monoclonal antibody to X protein [23], followed by analysis of the samples by SDS-PAGE, and finally by Western blotting using anti-HBs peptide antisera [6]. The specificity of both the monoclonal [23] and peptide antibodies [5, 22] used here was previously documented using recombinant DNA-derived HBxAg polypeptides (see below). Solid-phase assays were also used for detection of anti-HBx and anti-pol [4, 7]. In these assays, the appropriate X or polymerase synthetic peptides were used for coating microtiter wells. Test serum samples were then added to wells, followed by an enzyme-conjugated anti-immunoglobulin reagent and, finally, substrate for color development. Positive values were scored as being>2 SD above the mean of the negative controls, which again was a panel of human serum samples from uninfected individuals. The specificity of binding was determined by repeating the assays in the absence of the coating antigens and by preincubation of the test serum samples with excess soluble synthetic pep tides before assay [4, 7]. Specificity in the anti-HBx solid-phase assay was verified by determining whether serum samples could immunoprecipitate HBxAg+ samples from infectious serum or from recombinant DNA-generated HBxAg made in Escherichia coli. Briefly, each mixture of test serum and HBxAg+ sample was incubated overnight at 4°C, immunoprecipitated by the addition of protein A from Staphylococcus aureus (Boehringer Mannheim, Indianapolis), and analyzed by SDS-PAGE and by Western blotting using
1105
Horiike et al.
1106
JID I 99l;l64 ( December)
Table 1. Serologic profile of 85 hepatitis B virus (HBV)-infected patients. Positive for Diagnosis (n) Fulminant hepatitis (4) Acute hepatitis (7) Asymptomatic carrier ( 10) Chronic hepatitis (31 ) Liver cirrhosis (14) Primary hepatocellular carcinoma ( I9)
HBsAg 2 7 to 31 14 19
HBeAg anti-HBe HBxAg HBV DNA 0 4 5 20 5 5
4 3 5 II
0 1 0 8
8 12
I
0 4 4 16 4
2
I
anti-HBx anti-pol I
1 2 15 3 8
4 5 2 14 1 I
NOTE. HBsAg, HBeAg, HBxAg: HBV surface, e, and X antigens; anti-Hlle, anti-HBx, anti-pol: antibodies against HBeAg. HBxAg, and HBV polymerase.
Results HBxAg, anti-HBx. and anti-pol in patients with fulminant and acute hepatitis. HBxAg, anti-HBx, and anti-pol were determined in 85 HBV-infected Japanese patients. The results obtained were compared with the diagnosis and other serologic markers of HBV infection (table I). Among the 4 patients with fulminant hepatitis, 2 were HBsAg+, and all 4 were anti-HBe+ [29], but none of these individuals had detectable markers of viral replication in serum (i.e., HBeAg or HBV DNA). None of these patients had detectable anti-HD, demonstrating that the fulminant hepatitis was probably not due to infection with HDV [30]. However, all were anti-pol", suggesting that HBV replication occurs in patients with fulminant hepatitis in the absence of other serologic markers of HBV replication. All 7 patients with acute hepatitis were HBsAg+; 4 were HBV DNA+ and HBeAg+ (57%) and I of these was also HBxAg+ (14%) (table I). Two patients were anti-HBe+ in all sera tested, another developed anti-HBe, and yet another developed anti-HBx during the period of serum collection. Since anti-HBx and anti-Hlse are usually associated with decreased levels of detectable virus in serum [3, 7], it was of interest that 5 of the 7 patients with acute hepatitis (71 %) were anti-pol", Significantly, in serial sera collected from patients with fulminant and acute hepatitis, anti-pol was detected in 7 (64%) of II at the florid stage of hepatitis but in only 2 (18%) at the convalescent stage (P = .024). Together, these results suggest that anti-pol is a common, early marker
of HBV infection, as documented earlier [4], and that it could appear whether or not other markers of viral replication were detectable in serum. In contrast, HBxAg and antiHBx are not markers frequently detected in these types of HBV infections. HBxAg, anti-Hllx. and anti-pol in patients with chronic HBV infection. There was less evidence ofHBV replication in chronically infected patients than in those with acute and fulminant hepatitis (table I). In the 10 asymptomatic carriers tested, for example, only 5 (50%) were HBeAg+, and 4 of these were HBV DNA+. None of the asymptomatic carriers had detectable HBxAg. Among the HBeAg+ patients, 2 had anti-pol, while among the anti-HBe+ patients, 2 also had anti-HBx. Both patients with anti-pol were HBV DNA+ by dot-blot hybridization, while the 2 other patients with detectable anti-HBx were negative for HBV DNA in this assay. Similarly, patients with liver cirrhosis who had low frequencies of markers indicative ofviral replication also had a correspondingly low frequency of anti-pol (table I). Of 5 (36%) of 14 cirrhotic patients who were HBeAg+, none was anti-pol". These patients included 4 who were HBV DNA+, 1 who was HBxAg+, and another who was anti-HBx+. Among 8 other patients with anti-Hlse (57%), I had detectable anti-pol. Eight carriers with cirrhosis were anti-HBe+, 2 who had antiHBx and 1 with detectable anti-pol. Low frequencies of HBxAg, anti-HBx, and anti-pol were also observed in patients with hepatocellular carcinoma who had low frequencies of markers associated with viral replication (table I). HBxAg, anti-Hllx. and anti-pol in patients ~ith chronic hepatitis. Patients with chronic hepatitis B had higher frequencies ofHBxAg, anti-HBx, anti-pol, and other markers ofviral replication than did other chronically infected carriers studied (table 1). Of the 31 cases with chronic hepatitis, 20 (65%) were HBeAg+; of these, 16 (80%) were also HBV DNA+, 8 (40%) were HBxAg+, 10 (50%) were anti-HBx+, and 10 (50%) had detectable anti-pol. Patients in this category had undetectable levels of anti-HD, suggesting that the elevated serum transaminases observed in many cases was not due to HDV superinfection (data not shown). In contrast, among the II anti-HBe+ patients with chronic hepatitis, none had
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
ing signal by preincubation of the peptide antisera with the peptide 99 (l00 JLg/slice) for I h at 37°C. Analogous controls were obtained using antibodies generated against recombinant DNA-produced HBxAg by substituting the primary antibody with normal rabbit serum or by blocking reactivity with recombinant DNA-produced HBxAg polypeptide [22]. Statistical analysis. P values were calculated using Fisher's exact test for independence in a 2 X 2 table. The relationship between two compared characteristics was considered significant when P < .05.
JID 1991;164 (December)
New Markers in HBV Infection
c
B
A 12345678910111213
1107
1 2 3 4
kDa
5 6 78
2
3456789 kDa
1713Figure 1. Specificity controls for solid-phase immunoassays. Verification by immunoprecipitation of: A, hepatitis B X antigen (HBxAg) in serum samples scoring HBxAg+ (lanes 1-7) and HBxAg- (lanes 8-13) by solid-phase immunoassay; B, antibody to HBxAg (anti-HBx) in serum samples scoring anti-HBx+ (lanes 1-4) or anti-Hllx" (lanes 5-8) by solid-phase immunoassay (samples were mixed with HBxAg+ samples from infectious serum [odd-numbered lanes] or with recombinant DNA-generated HBxAg [even-numbered lanes]); C, polymerase polypeptide fragment (made by in vitro translation) immunoprecipitated by serum samples scoring positive (lanes 1-4) and negative (lanes 5-10) for antibody to the viral polymerase by solid-phase immunoassay.
detectable HBV DNA or HBxAg and only 4 (36%) were antipol", but 5 (45%) had detectable anti-HBx. Combining the results from both HBeAg+ and anti-Hlle" patients with chronic hepatitis, 10 (71 %) of the 14 cases with detectable anti-pol were also HBeAg+, while the remaining 4 were antiHBe+. All 8 patients with HBxAg were also HBeAg+, but anti-HBx was found in roughly half the patients regardless of HBeAg or anti-HBe status. Comparison of HBxAg, antiHBx, and anti-pol with other markers among patients with chronic hepatitis indicated a significant relationship between anti-pol and markers of HBV replication (HBeAg and HBV DNA) in serum (P = .03) as well as between HBxAg and these same markers (P = .013). Anti-HBx was not related to either HBeAg, HBV DNA, or anti-HBe (P> .05). The specificity of the serologic assays was determined using a variety of controls, as described previously [4-7]. HBxAg detection from serum in solid-phase assays was confirmed by immunoprecipitation of serum samples with a monoclonal antibody to X protein [23], followed by analysis of the sample by SDS-PAGE and finally by Western blotting using a mixture of well-characterized peptide antibodies against selected HBxAg sequences [5-7] (figure 1). Verification of anti-HBx and anti-pol in human serum samples was
carried out in immunoprecipitation experiments. These experiments confirmed the results of the solid-phase assays for these antibody specificities in the sera tested (figure 1). Characteristics of HBxAg in livers ofpatients with different types of HBV-associated liver disease. HBxAg was found in livers from HBV -infected patients with fulminant, acute, and chronic hepatitis, as well as from those with cirrhosis and hepatocellular carcinoma (table 2). Among patients with fulminant hepatitis, HBxAg was the only viral antigen detectable in one of three examined. In contrast, many patients with either acute or chronic hepatitis had detectable HBsAg, HBcAg, or both. In patients with chronic hepatitis and cirrhosis, HBsAg was the most frequent HBV antigen, followed by HBxAg and HBcAg. In all patients, most HBxAg staining in hepatocytes was cytoplasmic, although staining was also observed in the plasma membranes of hepatocytes in some cases (figure 2). Among 18 patients who had detectable HBxAg in liver, 13 (72%) had HBeAg in sera (P = .023) and II (61 %) had HBcAg in liver (P = .003). In contrast, among 11 patients who had no detectable HBxAg in liver, 3 had HBeAg in serum and 1 had HBcAg in liver. Hence, even though HBxAg was undetectable in the serum of many HBeAg+ patients, its presence in the liver still
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
-32
1108
Horiike et al.
Table 2.
Relationship of hepatitis B X antigen (HBxAg) in liver to other markers of hepatitis B virus infection. HBeAg in serum Diagnosis, HBxAg in liver (n)
NOTE.
*p
=
+
HBsAg in liver
+
0 0
I 2
0 0
I 0
0 I
I 0
0
9 3
0 2
8 I
I 4
9 5
0 0
0 0
2 I
0 0
2 I
2 I
0 0
3 0
2 2
3 0
2 2
3 2
2 0
13 3
5* 8
12 I
6t 10
15 9
3;
I
2
0 0
I 2
0 0
I
2
HBeAg. HBcAg, HBsAg: hepatitis B e, core. and surface antigen.
.023. tp = .0031, ;not significant compared with those HBxAg-.
correlated with ongoing viral replication. This tendency was especially pronounced in patients with chronic hepatitis. However, offive patients with hepatocellular carcinoma who had HBxAg in liver, two were negative for HBeAg in sera and HBcAg in liver. In addition, although two of three patients with liver cirrhosis had detectable HBxAg in tissue, none was HBcAg+. Further, many more chronically infected individuals were HBxAg+ in liver than were HBxAg+ in serum, even though the same primary antibodies were used to detect HBxAg in both circumstances, suggesting that HBxAg could be present in liver even when it was not detectable in serum. Comparison ofHBxAg and HBsAg in the liver showed no statistically significant relationship (table 2). Apparently, then, HBxAg in the liver closely correlates with markers of HBV replication, although it can be expressed in the liver in the absence of other markers, implying that it may play important roles in the pathogenesis of chronic infection. The specificity ofimmunohistochemical staining was demonstrated in a variety of control experiments. For example, HBxAg was completely absent from the liver tissue of eight patients with no evidence of HBV infection (data not shown). In addition, HBxAg staining was blocked by preincubation of the primary (peptide) antibody with an excess of the appropriate synthetic peptide used for immunization (figure 2). Normal rabbit serum did not show staining with any of the samples tested (data not shown). Sections stained with antibodies against recombinant DNA-expressed HBxAg
polypeptide resulted in patterns ofHBxAg staining similar to those observed with the peptide antibody. Specificity was again demonstrated by preincubation of the anti-recombinant HBxAg antibodies with an excess of recombinant DNA-generated HBxAg polypeptide before staining (data not shown). Additional negative controls, described previously [22], further showed the specificity of the antibodies used for immunochemical staining. The relationship between the HBxAg and anti-HBx system in serum, viral markers in the liver, and liver pathology is shown in table 3. All 8 patients who were HBxAg+ in serum also had elevated alanine aminotransferase levels (P = .029), as did 12 of 13 anti-Hllx" patients (P = .035), compared with those of patients who had neither HBxAg nor anti-HBx in serum. There was also a significant relationship between elevated transaminases and HBxAg in the liver (P = .035), suggesting that continued HBxAg expression closely correlates with biochemical evidence ofliver damage. Comparison of the HBxAg and anti-HBx markers in serum with viral replication (HBcAg) in the liver showed a significant relationship between HBcAg in the liver and serum HBxAg (P = .014) but not with anti-HBx (P> .05). There was also a statistically significant correlation between HBxAg in serum and HBeAg (P = .039), but not between anti-HBx and HBeAg (P > .05), suggesting a close correlation of HBxAg expression in both serum and liver with viral replication. Temporal association between the presence of HBxAg in liver and other markers ofinfection. Table 4 shows the time course of HBxAg, anti-HBx, and anti-pol in three HBV-infected patients. In a patient with acute hepatitis (case 3), HBxAg appeared at the same time as HBeAg and HBV DNA, "-'3 weeks after the appearance of HBsAg in the serum. This shows that HBxAg is an early marker of infection that accompanies other markers of viral replication in serum during acute infection. The results from immunohistochemical staining of a liver biopsy obtained from this patient several weeks before the rise in transaminases shows that HBsAg, HBcAg, and HBxAg can be expressed in the early stages of acute infection. Among patients with chronic hepatitis, one individual (case 39) seroconverted from HBxAg to anti-HBx in the course of 2-3 weeks. Although this patient remained persistently positive for both HBsAg and HBeAg, HBV DNA became undetectable by dot-blot hybridization during seroconversion, suggesting that the appearance of anti-HBx is accompanied by a decrease in or loss of virus from serum. Further, seroconversion from HBxAg to antiHBx was accompanied by a decrease in or loss of viral DNA from serum and a rise in transaminases. This pattern of events has also been seen in other populations [5-7] and suggests that this seroconversion may contribute to the removal of hepatocytes supporting viral replication. The immunohistochemical finding of HBsAg, HBcAg, and HBxAg both before and after seroconversion demon-
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
Fulminant hepatitis + (I) - (2) Acute hepatitis + (I) - (I) Chronic hepatitis + (9) - (5) Liver cirrhosis + (2) - (I) Primary hepatocellular carcinoma + (5) - (2) Total + (18) - (II)
+
HBcAg in liver
110 1991; 164 (December)
New Markers in HBV Infection
JID 1991;164 (December)
c
B
A
1109
strates that the appearance of anti-HBx may have limited effects upon the patterns of viral gene expression in the liver and that the persistent expression of one or more of these antigens in the liver continues to provide infected cells for destruction, as evidenced in this patient by another sharp rise in the transaminase levels in serum. In another patient with chronic hepatitis (case 44), the appearance of anti-HBx preceded the decreased levels of HBeAg in serum, the appearance ofanti-HBe, and the normalization oftransaminases, as documented in other populations [5-7]. In this case, normal-
Table 3. Relationship of hepatitis B X antigen (HBxAg) and antibody against HBxAg (anti-HBx) to alanine aminotransferase and hepatitis B core antigen (HBcAg) in livers of patients with chronic hepatitis. Alanine aminotransferase HBxAg/anti-HBx (n)
+/- (8) -/+ (13)
Characteristics of Hepatitis B X Antigen, Antibodies to X Antigen, and Antibodies to the Viral Polymerase during Hepatitis B Virus Infection Norio Horiike,* Baruch S. Blumberg, and Mark A. Feitelson*
Fox Chase Cancer Center, Philadelphia. Pennsylvania
Hepatitis B virus (HBV) infection is worldwide. Its pathogenesis is variable in that some individuals develop acute hepatitis while others develop the often deadly fulminant hepatitis and still others develop the asymptomatic chronic carrier state. Long-term infections could lead to the appearance of mild chronic persistent hepatitis or more severe chronic active hepatitis. Carriers with chronic active hepatitis are at high risk for the development of liver cirrhosis and primary hepatocellular carcinoma [I. 2]. The viral serum and liver markers associated with each of these disease states have been the subject of extensive investigations, but much remains to be done before a thorough understanding of their roles in the pathogenesis of liver disease is attained. The genome of HBV consists of four open-reading frames, each of which encodes one or more viral gene products. The first encodes a family of related hepatitis B surface antigen (HBsAg) polypeptides in the envelope of viral and subviral particles found in infectious serum. The second encodes the hepatitis B core antigen (HBcAg) polypeptides of the viral nucleocapsid. The third encodes products with reverse transcriptase, ribonuclease H, and DNA polymerase activities
Received 25 February 1991; revised 28 May 1991. Presented in part: Meeting on the Molecular Biology of Hepatitis B Viruses, University of California, San Diego, August 1990. Grant support: National Institutes of Health (RR-05895 and CA-40737, -06927, and -48656), Johnson and Johnson (Focused Giving grant 830702), Commonwealth of Pennsylvania, and Ehime University School of Medicine (to N.H.). Reprints or correspondence: Dr. Mark Feitelson, Department of Pathology and Cell Biology, Thomas Jefferson University School of Medicine. 1020 Locust sr., Philadelphia, PA 19107. * Present addresses: Third Department oflnternal Medicine, Ehime University School of Medicine. Ehime-ken, Japan (N.H.); Department of Pathology and Cell Biology, Thomas Jefferson University School of Medicine. Philadelphia (M.A. F.). The Journal ofInfectious Diseases 1991;164:1104-12 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6406-0009$01.00
required for viral replication. The fourth encodes hepatitis B X antigen (HBxAg) polypeptides [1, 3]. Recent work in renal dialysis patients who became transiently or persistently positive for HBsAg has shown that antibody directed against the viral polymerase (anti-pol) is the earliest marker of HBV infection [4], that HBxAg in the serum correlates with other markers of viral replication [5, 6], and that the appearance of antibodies against HBxAg (anti-HBx) correlates with a decrease in the amount ofdetectable virus in the serum [7]. The finding of fusion polypeptides with HBxAg plus core antigen determinants in viral nucleocapsid particles [8], in which viral replication is known to take place [9], further supports the inference that HBxAg is involved in replication. The finding of both HBxAg and polymerase polypeptides associated with viral replication suggests that they may be targets of antiviral immune responses likely to be important to the outcome of infection. This is especially important since liver disease is thought to be due to host immune responses against hepatocytes expressing one or more HBV antigens and not to a direct cytopathic effect of HBV [10]. If so, then the patterns of HBxAg [5, 6, 11-14], antibody to X protein [7, II, 1519], and anti-pol [4, 20, 21], which have only recently come under intense study, may be important markers of the hostvirus relationship. In this study we determined the presence, frequency, and distribution of these newly characterized HBV markers in sera and livers from Japanese patients infected with HBV. The results were compared with the clinical diagnosis to determine the role of these markers in the natural history of infection.
Materials and Methods Patients. Eighty-five adult HBV-infected Japanese patients recruited from the hospital at the Ehime University School of
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
The characteristics of hepatitis B virus (HBV) X antigen (HBxAg) and antibodies against the X antigen (anti-HBx) and the viral polymerase (anti-pol) were determined in 85 HBV-infected patients. HBxAg was detected in sera positive for HBV e antigen (HBeAg) and HBV DNA in patients with acute and chronic hepatitis, while anti-HBx appeared when markers of viral replication became undetectable. HBxAg was common in the liver among patients with chronic hepatitis independent of HBV replication markers but was closely correlated with elevated alanine aminotransferase, implying that HBxAg in liver may be important in the pathogenesis of chronic infection. Anti-pol was detected in many samples positive for HBeAg and HBV DNA and less often in serum samples without markers of HBV replication, suggesting that this marker could reflect ongoing viral replication in the liver, even though such markers were absent from sera.
110 1991; 164 (December)
New Markers in HBV Infection
the anti-HBx peptide antisera [6]. Recombinant HBxAg polypeptide in this case was made from the construct pET8c-X3 expressed in BL21(DE3)pLysS as described earlier [24], which was a gift ofW. Robinson (Stanford University, Stanford, CA). Specificity in the anti-pol solid-phase assay was verified by determining whether test sera could immunoprecipitate a recombinant DNA-derived polymerase polypeptide fragment. The immunoprecipitation was carried out exactly as described for anti-HBx above, except that the material for immunoprecipitation was derived from in vitro transcription and translation of a polymerase expression vector. The polymerase fragment chosen for expression included the 3' end of the polymerase gene from the full-length clone of HBV adw, [25], which was amplified using primers spanning nucleotide positions 2436-2412 (MF04) and 631-660 (MFHP 2 ) . These primers were made in the oligonucleotide synthesis facility at the Fox Chase Cancer Center. This region of the polymerase gene was chosen for expression because it carried imrnunodominant determinants recognized by most patients with an anti-pol response [4, 26]. The expression product (32 kDa) was identified by irnrnunoprecipitation using appropriate anti-pol peptide antisera [4]. Tritiumlabeled polymerase polypeptide was then used for immunoprecipitation followed by SDS-PAGE, and the results were visualized by fluorography [27]. Antibodies against hepatitis delta virus (anti-HD) were measured using a commercially available assay (Abbott). Serum alanine aminotransferase levels were measured by standard techniques [4]. Peptide synthesis and generation of corresponding antisera. Synthetic pep tides from the polymerase and X regions of HBV DNA (subtype ayw) [28] were made at the solid-phase peptide synthesis facility at the Fox Chase Cancer Center. Corresponding antisera were prepared as described [4, 5, 7]. Immunohistochemical study. The liver specimens were fixed in formalin and embedded in paraffin. Sections (6 JLm thick) were used for staining as described [22]. HBxAg was examined by an avidin-biotin complex method using a polyclonal rabbit antibody to an HBxAg-specific peptide spanning residues 100114 (anti-99) [7]. Briefly, slides were blocked with excess avidin (0.05%, 10 min) and biotin (0.05%, 10 min) and then incubated overnight with a 1:400 dilution ofanti-99. After washing, slides were incubated with biotinylated goat anti-rabbit immunoglobulin for 30 min, then with avidin-biotinylated peroxidase complex (Vectastain, ABC kit; Vector Laboratories, Burlingame, CA), and finally with diarninobenzidine substrate [22]. Slides were counterstained with hematoxylin and observed by light microscopy under code. Some slides were stained with a primary antibody raised against full-length, 17-kDa HBxAg polypeptide, made by recombinant DNA techniques and provided by W. Robinson. HBsAg and HBcAg were detected by using commercially available kits (Dako, Copenhagen) following the instructions of the manufacturer. HBxAg expression in the liver was graded as follows: -, negative; +, positive, with a spotty distribution in tissue; ++, positive, with a lobular or diffuse distribution in tissue. The specificity of staining for HBxAg was demonstrated by replacement of anti-99 with preimmune rabbit serum or by blocking the stain-
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
Medicine were studied. Among them, 4 had fulminant hepatitis, 7 had acute hepatitis, 10 were asymptomatic carriers (defined as persistently HBsAg+ patients lacking clinical and biochemical evidence of liver disease), 31 had chronic hepatitis, 14 had cirrhosis, and 19 had hepatocellular carcinoma. For patients from whom liver biopsy materials were available (3 with fulminant hepatitis, 2 with acute hepatitis, 14 with chronic hepatitis, 3 with cirrhosis, and 7 with hepatocellular carcinoma), paraffin-embedded sections were stained for HBsAg, HBcAg, and HBxAg as described [22]. HBV and hepatitis delta virus (HDV) markers and alanine aminotransferase levels in serum. The presence of serum HBsAg, antibodies against HBsAg (anti-Hlss), HBV e antigen (HBeAg), antibodies against HBeAg (anti-Hlse), and antibodies against HBcAg were determined using commercial radioimmunoassays (Abbott Laboratories, North Chicago, IL). HBV DNA was tested by dot hybridization using 1 JLl of each serum sample as described [5]. HBxAg was determined exactly as described [5]. Briefly, serum samples were initially screened for HBxAg by ELISA. Microtiter wells were coated with human serum for testing, then with anti-HBx made in rabbits against HBxAg synthetic peptides; finally, an enzyme-conjugated anti-immunoglobulin and substrate for color development were added. Positive values were scored as being >2 SD above the mean of the negative controls, which was a panel of human serum samples from uninfected individuals. The specificity of binding was determined by repeating the assay with normal rabbit serum in the place of peptide antibodies. Specificity was also shown by preincubation of the peptide antibodies with the immunizing peptides before assay. Specificity of the solid-phase assay was verified by immunoprecipitation of selected positive and negative serum samples using a monoclonal antibody to X protein [23], followed by analysis of the samples by SDS-PAGE, and finally by Western blotting using anti-HBs peptide antisera [6]. The specificity of both the monoclonal [23] and peptide antibodies [5, 22] used here was previously documented using recombinant DNA-derived HBxAg polypeptides (see below). Solid-phase assays were also used for detection of anti-HBx and anti-pol [4, 7]. In these assays, the appropriate X or polymerase synthetic peptides were used for coating microtiter wells. Test serum samples were then added to wells, followed by an enzyme-conjugated anti-immunoglobulin reagent and, finally, substrate for color development. Positive values were scored as being>2 SD above the mean of the negative controls, which again was a panel of human serum samples from uninfected individuals. The specificity of binding was determined by repeating the assays in the absence of the coating antigens and by preincubation of the test serum samples with excess soluble synthetic pep tides before assay [4, 7]. Specificity in the anti-HBx solid-phase assay was verified by determining whether serum samples could immunoprecipitate HBxAg+ samples from infectious serum or from recombinant DNA-generated HBxAg made in Escherichia coli. Briefly, each mixture of test serum and HBxAg+ sample was incubated overnight at 4°C, immunoprecipitated by the addition of protein A from Staphylococcus aureus (Boehringer Mannheim, Indianapolis), and analyzed by SDS-PAGE and by Western blotting using
1105
Horiike et al.
1106
JID I 99l;l64 ( December)
Table 1. Serologic profile of 85 hepatitis B virus (HBV)-infected patients. Positive for Diagnosis (n) Fulminant hepatitis (4) Acute hepatitis (7) Asymptomatic carrier ( 10) Chronic hepatitis (31 ) Liver cirrhosis (14) Primary hepatocellular carcinoma ( I9)
HBsAg 2 7 to 31 14 19
HBeAg anti-HBe HBxAg HBV DNA 0 4 5 20 5 5
4 3 5 II
0 1 0 8
8 12
I
0 4 4 16 4
2
I
anti-HBx anti-pol I
1 2 15 3 8
4 5 2 14 1 I
NOTE. HBsAg, HBeAg, HBxAg: HBV surface, e, and X antigens; anti-Hlle, anti-HBx, anti-pol: antibodies against HBeAg. HBxAg, and HBV polymerase.
Results HBxAg, anti-HBx. and anti-pol in patients with fulminant and acute hepatitis. HBxAg, anti-HBx, and anti-pol were determined in 85 HBV-infected Japanese patients. The results obtained were compared with the diagnosis and other serologic markers of HBV infection (table I). Among the 4 patients with fulminant hepatitis, 2 were HBsAg+, and all 4 were anti-HBe+ [29], but none of these individuals had detectable markers of viral replication in serum (i.e., HBeAg or HBV DNA). None of these patients had detectable anti-HD, demonstrating that the fulminant hepatitis was probably not due to infection with HDV [30]. However, all were anti-pol", suggesting that HBV replication occurs in patients with fulminant hepatitis in the absence of other serologic markers of HBV replication. All 7 patients with acute hepatitis were HBsAg+; 4 were HBV DNA+ and HBeAg+ (57%) and I of these was also HBxAg+ (14%) (table I). Two patients were anti-HBe+ in all sera tested, another developed anti-HBe, and yet another developed anti-HBx during the period of serum collection. Since anti-HBx and anti-Hlse are usually associated with decreased levels of detectable virus in serum [3, 7], it was of interest that 5 of the 7 patients with acute hepatitis (71 %) were anti-pol", Significantly, in serial sera collected from patients with fulminant and acute hepatitis, anti-pol was detected in 7 (64%) of II at the florid stage of hepatitis but in only 2 (18%) at the convalescent stage (P = .024). Together, these results suggest that anti-pol is a common, early marker
of HBV infection, as documented earlier [4], and that it could appear whether or not other markers of viral replication were detectable in serum. In contrast, HBxAg and antiHBx are not markers frequently detected in these types of HBV infections. HBxAg, anti-Hllx. and anti-pol in patients with chronic HBV infection. There was less evidence ofHBV replication in chronically infected patients than in those with acute and fulminant hepatitis (table I). In the 10 asymptomatic carriers tested, for example, only 5 (50%) were HBeAg+, and 4 of these were HBV DNA+. None of the asymptomatic carriers had detectable HBxAg. Among the HBeAg+ patients, 2 had anti-pol, while among the anti-HBe+ patients, 2 also had anti-HBx. Both patients with anti-pol were HBV DNA+ by dot-blot hybridization, while the 2 other patients with detectable anti-HBx were negative for HBV DNA in this assay. Similarly, patients with liver cirrhosis who had low frequencies of markers indicative ofviral replication also had a correspondingly low frequency of anti-pol (table I). Of 5 (36%) of 14 cirrhotic patients who were HBeAg+, none was anti-pol". These patients included 4 who were HBV DNA+, 1 who was HBxAg+, and another who was anti-HBx+. Among 8 other patients with anti-Hlse (57%), I had detectable anti-pol. Eight carriers with cirrhosis were anti-HBe+, 2 who had antiHBx and 1 with detectable anti-pol. Low frequencies of HBxAg, anti-HBx, and anti-pol were also observed in patients with hepatocellular carcinoma who had low frequencies of markers associated with viral replication (table I). HBxAg, anti-Hllx. and anti-pol in patients ~ith chronic hepatitis. Patients with chronic hepatitis B had higher frequencies ofHBxAg, anti-HBx, anti-pol, and other markers ofviral replication than did other chronically infected carriers studied (table 1). Of the 31 cases with chronic hepatitis, 20 (65%) were HBeAg+; of these, 16 (80%) were also HBV DNA+, 8 (40%) were HBxAg+, 10 (50%) were anti-HBx+, and 10 (50%) had detectable anti-pol. Patients in this category had undetectable levels of anti-HD, suggesting that the elevated serum transaminases observed in many cases was not due to HDV superinfection (data not shown). In contrast, among the II anti-HBe+ patients with chronic hepatitis, none had
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
ing signal by preincubation of the peptide antisera with the peptide 99 (l00 JLg/slice) for I h at 37°C. Analogous controls were obtained using antibodies generated against recombinant DNA-produced HBxAg by substituting the primary antibody with normal rabbit serum or by blocking reactivity with recombinant DNA-produced HBxAg polypeptide [22]. Statistical analysis. P values were calculated using Fisher's exact test for independence in a 2 X 2 table. The relationship between two compared characteristics was considered significant when P < .05.
JID 1991;164 (December)
New Markers in HBV Infection
c
B
A 12345678910111213
1107
1 2 3 4
kDa
5 6 78
2
3456789 kDa
1713Figure 1. Specificity controls for solid-phase immunoassays. Verification by immunoprecipitation of: A, hepatitis B X antigen (HBxAg) in serum samples scoring HBxAg+ (lanes 1-7) and HBxAg- (lanes 8-13) by solid-phase immunoassay; B, antibody to HBxAg (anti-HBx) in serum samples scoring anti-HBx+ (lanes 1-4) or anti-Hllx" (lanes 5-8) by solid-phase immunoassay (samples were mixed with HBxAg+ samples from infectious serum [odd-numbered lanes] or with recombinant DNA-generated HBxAg [even-numbered lanes]); C, polymerase polypeptide fragment (made by in vitro translation) immunoprecipitated by serum samples scoring positive (lanes 1-4) and negative (lanes 5-10) for antibody to the viral polymerase by solid-phase immunoassay.
detectable HBV DNA or HBxAg and only 4 (36%) were antipol", but 5 (45%) had detectable anti-HBx. Combining the results from both HBeAg+ and anti-Hlle" patients with chronic hepatitis, 10 (71 %) of the 14 cases with detectable anti-pol were also HBeAg+, while the remaining 4 were antiHBe+. All 8 patients with HBxAg were also HBeAg+, but anti-HBx was found in roughly half the patients regardless of HBeAg or anti-HBe status. Comparison of HBxAg, antiHBx, and anti-pol with other markers among patients with chronic hepatitis indicated a significant relationship between anti-pol and markers of HBV replication (HBeAg and HBV DNA) in serum (P = .03) as well as between HBxAg and these same markers (P = .013). Anti-HBx was not related to either HBeAg, HBV DNA, or anti-HBe (P> .05). The specificity of the serologic assays was determined using a variety of controls, as described previously [4-7]. HBxAg detection from serum in solid-phase assays was confirmed by immunoprecipitation of serum samples with a monoclonal antibody to X protein [23], followed by analysis of the sample by SDS-PAGE and finally by Western blotting using a mixture of well-characterized peptide antibodies against selected HBxAg sequences [5-7] (figure 1). Verification of anti-HBx and anti-pol in human serum samples was
carried out in immunoprecipitation experiments. These experiments confirmed the results of the solid-phase assays for these antibody specificities in the sera tested (figure 1). Characteristics of HBxAg in livers ofpatients with different types of HBV-associated liver disease. HBxAg was found in livers from HBV -infected patients with fulminant, acute, and chronic hepatitis, as well as from those with cirrhosis and hepatocellular carcinoma (table 2). Among patients with fulminant hepatitis, HBxAg was the only viral antigen detectable in one of three examined. In contrast, many patients with either acute or chronic hepatitis had detectable HBsAg, HBcAg, or both. In patients with chronic hepatitis and cirrhosis, HBsAg was the most frequent HBV antigen, followed by HBxAg and HBcAg. In all patients, most HBxAg staining in hepatocytes was cytoplasmic, although staining was also observed in the plasma membranes of hepatocytes in some cases (figure 2). Among 18 patients who had detectable HBxAg in liver, 13 (72%) had HBeAg in sera (P = .023) and II (61 %) had HBcAg in liver (P = .003). In contrast, among 11 patients who had no detectable HBxAg in liver, 3 had HBeAg in serum and 1 had HBcAg in liver. Hence, even though HBxAg was undetectable in the serum of many HBeAg+ patients, its presence in the liver still
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
-32
1108
Horiike et al.
Table 2.
Relationship of hepatitis B X antigen (HBxAg) in liver to other markers of hepatitis B virus infection. HBeAg in serum Diagnosis, HBxAg in liver (n)
NOTE.
*p
=
+
HBsAg in liver
+
0 0
I 2
0 0
I 0
0 I
I 0
0
9 3
0 2
8 I
I 4
9 5
0 0
0 0
2 I
0 0
2 I
2 I
0 0
3 0
2 2
3 0
2 2
3 2
2 0
13 3
5* 8
12 I
6t 10
15 9
3;
I
2
0 0
I 2
0 0
I
2
HBeAg. HBcAg, HBsAg: hepatitis B e, core. and surface antigen.
.023. tp = .0031, ;not significant compared with those HBxAg-.
correlated with ongoing viral replication. This tendency was especially pronounced in patients with chronic hepatitis. However, offive patients with hepatocellular carcinoma who had HBxAg in liver, two were negative for HBeAg in sera and HBcAg in liver. In addition, although two of three patients with liver cirrhosis had detectable HBxAg in tissue, none was HBcAg+. Further, many more chronically infected individuals were HBxAg+ in liver than were HBxAg+ in serum, even though the same primary antibodies were used to detect HBxAg in both circumstances, suggesting that HBxAg could be present in liver even when it was not detectable in serum. Comparison ofHBxAg and HBsAg in the liver showed no statistically significant relationship (table 2). Apparently, then, HBxAg in the liver closely correlates with markers of HBV replication, although it can be expressed in the liver in the absence of other markers, implying that it may play important roles in the pathogenesis of chronic infection. The specificity ofimmunohistochemical staining was demonstrated in a variety of control experiments. For example, HBxAg was completely absent from the liver tissue of eight patients with no evidence of HBV infection (data not shown). In addition, HBxAg staining was blocked by preincubation of the primary (peptide) antibody with an excess of the appropriate synthetic peptide used for immunization (figure 2). Normal rabbit serum did not show staining with any of the samples tested (data not shown). Sections stained with antibodies against recombinant DNA-expressed HBxAg
polypeptide resulted in patterns ofHBxAg staining similar to those observed with the peptide antibody. Specificity was again demonstrated by preincubation of the anti-recombinant HBxAg antibodies with an excess of recombinant DNA-generated HBxAg polypeptide before staining (data not shown). Additional negative controls, described previously [22], further showed the specificity of the antibodies used for immunochemical staining. The relationship between the HBxAg and anti-HBx system in serum, viral markers in the liver, and liver pathology is shown in table 3. All 8 patients who were HBxAg+ in serum also had elevated alanine aminotransferase levels (P = .029), as did 12 of 13 anti-Hllx" patients (P = .035), compared with those of patients who had neither HBxAg nor anti-HBx in serum. There was also a significant relationship between elevated transaminases and HBxAg in the liver (P = .035), suggesting that continued HBxAg expression closely correlates with biochemical evidence ofliver damage. Comparison of the HBxAg and anti-HBx markers in serum with viral replication (HBcAg) in the liver showed a significant relationship between HBcAg in the liver and serum HBxAg (P = .014) but not with anti-HBx (P> .05). There was also a statistically significant correlation between HBxAg in serum and HBeAg (P = .039), but not between anti-HBx and HBeAg (P > .05), suggesting a close correlation of HBxAg expression in both serum and liver with viral replication. Temporal association between the presence of HBxAg in liver and other markers ofinfection. Table 4 shows the time course of HBxAg, anti-HBx, and anti-pol in three HBV-infected patients. In a patient with acute hepatitis (case 3), HBxAg appeared at the same time as HBeAg and HBV DNA, "-'3 weeks after the appearance of HBsAg in the serum. This shows that HBxAg is an early marker of infection that accompanies other markers of viral replication in serum during acute infection. The results from immunohistochemical staining of a liver biopsy obtained from this patient several weeks before the rise in transaminases shows that HBsAg, HBcAg, and HBxAg can be expressed in the early stages of acute infection. Among patients with chronic hepatitis, one individual (case 39) seroconverted from HBxAg to anti-HBx in the course of 2-3 weeks. Although this patient remained persistently positive for both HBsAg and HBeAg, HBV DNA became undetectable by dot-blot hybridization during seroconversion, suggesting that the appearance of anti-HBx is accompanied by a decrease in or loss of virus from serum. Further, seroconversion from HBxAg to antiHBx was accompanied by a decrease in or loss of viral DNA from serum and a rise in transaminases. This pattern of events has also been seen in other populations [5-7] and suggests that this seroconversion may contribute to the removal of hepatocytes supporting viral replication. The immunohistochemical finding of HBsAg, HBcAg, and HBxAg both before and after seroconversion demon-
Downloaded from http://jid.oxfordjournals.org/ at New York University on September 14, 2015
Fulminant hepatitis + (I) - (2) Acute hepatitis + (I) - (I) Chronic hepatitis + (9) - (5) Liver cirrhosis + (2) - (I) Primary hepatocellular carcinoma + (5) - (2) Total + (18) - (II)
+
HBcAg in liver
110 1991; 164 (December)
New Markers in HBV Infection
JID 1991;164 (December)
c
B
A
1109
strates that the appearance of anti-HBx may have limited effects upon the patterns of viral gene expression in the liver and that the persistent expression of one or more of these antigens in the liver continues to provide infected cells for destruction, as evidenced in this patient by another sharp rise in the transaminase levels in serum. In another patient with chronic hepatitis (case 44), the appearance of anti-HBx preceded the decreased levels of HBeAg in serum, the appearance ofanti-HBe, and the normalization oftransaminases, as documented in other populations [5-7]. In this case, normal-
Table 3. Relationship of hepatitis B X antigen (HBxAg) and antibody against HBxAg (anti-HBx) to alanine aminotransferase and hepatitis B core antigen (HBcAg) in livers of patients with chronic hepatitis. Alanine aminotransferase HBxAg/anti-HBx (n)
+/- (8) -/+ (13)
